Was there a discussion I missed somewhere on making sequential uuid's
the default algorithm? We decided against it when applying the patch
from COUCHDB-465 [1] and I still hesitate to use anything other than
completely random as the default. With the sequential algorithm we're
leaking doc creation times in a non-obvious manner.

I'd rather just see a wiki page on something like "Configuring CouchDB
for improved performance" that went over any config settings to speed
things up with relevant explanations for each. This could also cover
batch inserts with the logic behind the changes.

Paul Davis

[1] https://issues.apache.org/jira/browse/COUCHDB-465

On Sat, Jan 9, 2010 at 1:23 PM,  <[email protected]> wrote:
> Author: jchris
> Date: Sat Jan  9 18:23:16 2010
> New Revision: 897509
>
> URL: http://svn.apache.org/viewvc?rev=897509&view=rev
> Log:
> make sequential uuids the default
>
> Modified:
>    couchdb/trunk/etc/couchdb/default.ini.tpl.in
>
> Modified: couchdb/trunk/etc/couchdb/default.ini.tpl.in
> URL: 
> http://svn.apache.org/viewvc/couchdb/trunk/etc/couchdb/default.ini.tpl.in?rev=897509&r1=897508&r2=897509&view=diff
> ==============================================================================
> --- couchdb/trunk/etc/couchdb/default.ini.tpl.in (original)
> +++ couchdb/trunk/etc/couchdb/default.ini.tpl.in Sat Jan  9 18:23:16 2010
> @@ -99,7 +99,7 @@
>  ;     random prefix is regenerated and the process starts over.
>  ;   utc_random - Time since Jan 1, 1970 UTC with microseconds
>  ;     First 14 characters are the time in hex. Last 18 are random.
> -algorithm = random
> +algorithm = sequential
>
>  [stats]
>  ; rate is in milliseconds
>
>
>

Reply via email to